Reagents and processes currently used for protein cross-linking generally depend upon the incorporation the cross-linking reagent into the protein matrix to cross-link the .epsilon.-amino groups of lysine, hydroxylysine, and/or other groups in the protein. Common cross-linking reagents in such processes include formaldehyde and glutaraldehyde; other processes include the introduction of a phthaloyl or adipoyl moiety into the protein via phthaloyl dichloride adipoyl dichloride , respectively, and/or the introduction of a mercaptan for oxidization to a disulfide bond.
The cross-linking processes, reactions and reagents the prior art vary, but most involve incorporating the reagent into or around the protein. For example, recent data by Cheung and Nimni (Connec. Tissue Res. 10:201 (1982) and Connec. Tissue Res. 13:109 (1984)) on the cross-linking reagent glutaraldehyde indicate that when this reagent is used to treat collagen fibrils, for example, a polymeric-like coating forms around the fibrils, resulting in stiffer collagen matrix.
In contrast, the cross-linking method disclosed and claimed herein does not depend upon the incorporation of a cross-linking reagent into the material to be cross-linked or the coating of the material with a cross-linked reagent. The present process involves the use of a photooxidative dye which acts as a cross-linking oxidation catalyst or promotor and which can be removed from the cross-linked product.
The use of photooxidative catalysts in various photooxidation processes has been previously reported (see e.g. , Ray, Method in Enzymol. 11:490 (1967); Westhead, Blochem 4:10 (1965); Ray and Koshland, Jr. , J. Biological Chem. 18:409 (1967); and Foote, Science 162:3857 (1968) . However, they do not appear to have been used for cross-linking proteinaceous materials. For instance, Ray and Koshland, Jr. , supra, used methylene blue and Light to photooxidize the enzyme phosphoglucomutase in an attempt to identify the amino acid residues of that protein which are essential to the activity of the enzyme by selective destruction of amino acids. Likewise, Westhead, supra, inactivated yeast enolase by photooxidation of histidine residues with the dye rose bengal.
Excitation of a dye by light has also been used to covalently couple the dye to a protein (Brandt, et al., Biochemistry 13: 4758 (1974)), and that technique has led to a method of dye-sensitized photolabeling of proteins (Brantit, et al. , Anal. Blochem. 93:601 (1980) . Although the technique is useful for such purposes as the study of the molecular arrangement of proteinaceous membrane components ( Id. ) and protein conformation (Hemmendorff, et al., Blochem. Biophys. Acta 667: 15 (1981)) , the technique does not appear to introduce inter- and/or intra-molecular cross-links into the protein matrix.
A dye-catalyzed process said to be useful for preparing thermostable, irreversibly cross-linked collagenous polymers is described in U.S. Pat. No. 3,152,976. This patent alleges that the product resulting from that process is characterized by certain physical-chemical properties similar to those obtained by prior art tanning processes. However, the subsequent data presented in that patent do not support a conclusion that the product of that process possesses the properties of products of prior art tanning processes which would make that product a useful biomaterial for such applications as vascular grafts, heart valves, pericardial patches, injectable collagen, or replacement ligaments or tendoils. Instead, that reference states that the product is more susceptible to enzymatic degradation than "uncross-linked" collagen. Such results are, of course, totally contrary to the use of such a product as, for instance, a heart valve (imagine a heart valve digested by even tile mildly proteolytic enzyme papain in hours, or even seconds, as described in Example VII of that reference). These seemingly anomalous results can perhaps in part be explained by the apparent motivation for making the invention described in that patent, namely the formation of "shaped articles" such as sponges or fibrils (sutures?), ostensibly of a type which can be implanted in the body without type need for subsequent removal.
The results reported in the '976 patent can perhaps also be explained by a close examination of the process described therein. For instance, the reference describes the preparation of a "starting material" on which the process set out in that patent is conducted by dispersing collagenous material in aqueous acid solution. Acid has the well-known effect of denaturing the protein comprising the collagen fibril. It is, of course, the three-dimensional structure of the proteins comprising the collagen fibril which imparts to the fibril the unique properties of collagen; change that structure and the protein cannot interact in the manner needed to give rise to those properties. A further explanation for the results described in that patent is suggested by P. H. von Hipple, "Structural and Stabilization of the Collagen Molecule in Solution" (in Treatise on Collagen, Vol. 1: Chemistry of Collagen, G. N. Ramachandran (Ed.), London: Academic Press Inc. (London) Ltd. (1967), pp. 253-338 at 262), reporting that collagen molecules extracted by acid and neutral salt procedures differ in the extent to which they are covalently cross-linked, size, shape, interaction properties and rate of fiber formation. Although based on preliminary data such that the author was careful to point out that results had been reported by other investigators which did not show any differences, subsequent experimentation supports the existence of such differences.
In light of this prior art, it was surprising to find that photooxidation of a protein in the presence of a photo-catalyst and sufficient oxygen, under controlled conditions of pH and temperature, cross-linked and stabilized the collagen to, for instance, enzymatic degradation, without stiffening the matrix like in conventional tanning processes.